Easter Term 2012

Tuesday 15th May 2012, 14:30-15:30
LARGE Seminar Room, 1st Floor, Institute of Public Health

Jonathan Bartlett
London School Of Hygiene amd Tropical Medicine

“Congenial multiple imputation of partially observed covariates within the full conditional specification framework”

Summary: Missing covariate data is a common issue in epidemiological and clinical research, and is often dealt with using multiple imputation (MI). When the analysis model is non-linear, or contains non-linear (e.g. squared) or interaction terms, this complicates the imputation of covariates. Standard software implementations of MI typically impute covariates from models that are uncongenial with such analysis models. We show how imputation by full conditional specification, a popular approach for performing MI, can be modified so that covariates are imputed from a model which is congenial with the analysis model. We investigate through simulation the performance of this proposal, and compare it to passive imputation of non-linear or interaction terms and the `just another variable’ approach. Our proposed approach provides consistent estimates provided the imputation models and analysis models are correctly specified and data are missing at random. In contrast, passive imputation of non-linear or interaction terms generally results in inconsistent estimates of the parameters of the model of interest, while the `just another variable' approach gives consistent results only for linear models and only if data are missing completely at random. Furthermore, simulation results suggest that even under imputation model mis-specification our proposed approach gives estimates which are substantially less biased than estimates based on passive imputation. The proposed approach is illustrated using data from the National Child Development Survey in which the analysis model contains both non-linear and interaction terms.

Tuesday 19th June 2012, 11:00-12:00
LARGE Seminar Room, 1st Floor, Institute of Public Health

Lisa Hampson
Lancaster University

“Optimal data combination in seamless Phase II/III clinical trials”

Summary: We consider seamless Phase II/III clinical trials which compare K treatments against a common control in stage 1 and select the most promising for further testing against control in stage 2. Such a trial requires careful upfront planning if it is to win regulatory acceptance as a pivotal study. For seamless trials to be attractive, this increased planning should be offset by efficiency gains made possible because data accumulated across the study are combined to make a final decision on the efficacy of the selected treatment. We derive optimal versions of final decision rules maximising power. This is a multivariate decision problem because properties of rules depend on a vector of means.

Rules with the correct familywise error rate maximising power for different configurations of means are found as solutions to Bayes decision problems. Different solutions are found as the shape of the mean vector changes but we find only small gains in power are possible by making strong assumptions about the structure of the mean vector. By studying procedures with optimal decision rules, we assess the efficiency of alternative proposals, namely closed testing procedures based on p-value combination rules, and rules using only data on the selected treatment and control for final decisions. For procedures with efficient decision rules, we find that Phase II observations on the selected treatment and control retain between 22-98% of their value as Phase III observations. Thus, efficient seamless designs can offer large savings in sample size which may have important implications, for example, for the feasibility of trials in rare diseases.